Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - Roderick Klein

Pages: [1] 2 3 ... 44
1
Programming / Re: DosDevIOCtl disabling keyboard in VIO mode
« on: March 07, 2024, 12:23:01 am »
But I also do not understand why you would be dogmatic about using 32-bit only.
Firstly from a technical cleanliness perspective. I thought that OS/2 2.0 was about upgrading to 32-bit. 30 years later and I'm still required to deal with 16-bit? Not what I want to do. Especially not for one problematic function (DosDevIOCtl) that would force me into 16-bit. I'd rather be forced into fullscreen.

Secondly, the pure 32-bit allows me to write a mini-clone of OS/2 either as an app that runs under another OS (e.g. UC386L that runs under Linux) or as a standalone OS (PDOS/386). Both available from http://pdos.org - new version that implements DosDevIOCtl uploaded a couple of minutes ago.

Note that "mini clone" means that it runs *certain* OS/2 executables - basically just the ones that were built with PDPCLIB (see makefile.wat). So anything that is C90-compliant, or uses ANSI x3.64 "extensions" and has been built with PDPCLIB or equivalent (I doubt that there is anything equivalent currently, but there is no barrier to that).

Quote
OS/2 has a strong 16-bit heritage, that's just the way it is.
And it doesn't need to be. With proper funding, OS/2 could be stripped internally of any 16-bit vestiges that affect a pure 32-bit application. As well as being a commercial competitor to Windows (note that I am similarly able to convert OS/2 into a mini Win32 clone with the use of win32os2 at pdos.org using the same technique). And yes, I know that Odin for OS/2 and Linux WINE are nominally superior (but neither are what I want).

So anyway, I want to be able to build a 32-bit OS/2 binary today, that runs on all those environments. And even if I lose the source code, the binary remains pure as better OS/2 clones come online.

Having worked on OS/2 for 25 years and more then 10 years on eComStation project I have somewhat experience in this. The issue is that it seems the funding that would be needed never arrived. It seemed nobody was interested in this approach.  Could it be done, I guess so. OS/2 also had ODIN on it, so in the other direction that might work.  However it seems that in most cases it was cheaper and faster to create a VM. Take Virtual PC with OS/2, Svista done by Serenity Systems. I seem to under years ago a France OS/2 customer wanted to move froim OS/2 to Linux. The problem was they had lost the sources of the OS/2 app. I seem to understand IBM tried to create an API layer on Linux. The project was never finished.

Anyway I am not saying it can not be done. I have no clue how much such a project would cost.  But I am skeptical at this moment if it will ever come off the ground.

Roderick

2
Applications / Re: XWP v1.0.16
« on: March 04, 2024, 09:52:38 am »
Hello

What do you think it is needed for lSwitcher to also respect the XWP font sizes too?

I opened a ticket in case Gregg Young wants to take a look at it.
- http://trac.netlabs.org/lswitcher/ticket/81

Regards

Gregg is on vacation right now.  He is aware of this ISwitcher issue. He mentioned this issue at Warpstock.

Roderick

3
Storage / Re: Boot from iscsi or fiber channel
« on: February 12, 2024, 12:20:22 pm »
I would like to have a diskless ArcaOS (or Warp 4, but preferably ArcaOS) machine. For that I would like to boot and run diskless from iSCSI or Fiber Channel or FCoE (Fiberchannel over Ethernet.)

Either way it will be booting from a Netapp filter.

It's just much more convenient for us, as it allows us to take filesystem snapshots, replicate storage remotely; and it gets automatically backed up; plus naturally, it's a very reliable SAN underneath with parity; and finally, it's more "aligned" with how we run things around here.

The initial OS installation can be done to Virtualbox (because I don't think you can readily install to FC or iSCSI), then the contents of the disk *could* be moved (more precisely dd if=virtualboxdisk.raw of=/dev/remoteblockdevice bs=512 or something like that) to the SAN block device, but I'm not certain how things would boot (ECSBOOT.COM?); naturally, I could use a Qlogic HBA with iSCSI boot support (sometimes they do that for windows or AIX) but I'm not sure if that was ever attempted for OS/2; although, because it's a fairly common way to do things in large organizations i would be surprised to hear that it was never done with OS/2.

GOALS: Note that I don't want to boot or run diskless os/2 from os/2 (or ecomstation or another operating system) or run os/2 "virtualized" here; the goal is to boot and run os/2 from a machine with no disk. So please refrain asking "why would you do this", "use a raid controller", "use reliable drives", etc. Finally, the GOAL is not to simply boot, but have the block device exposed over network or fiber channel; or basically run "diskless" via iscsi/FCoE or fiber channel.

Quick search on Google didn't show anything extremely useful (which hardware to use, which HBAs are supported etc) or how to accomplish this, so I'm turning to the community for ideas and pointers.

Thanks in advance -- J.

Sorry best I can tell this is game over. Maybe just maybe the current ArcaOS UEFI loader MIGHT work to load from iSCSI. The problem is that would load ArcaOS for a small portion. But once ArcaOS starts to boot with the UEFI loader your dead in the water.
The problem is that you would need some sort of iSCSI add driver (I think). This is really uncharted terrain and I think your chances or low of getting this to work.


Roderick Klein

4
Setup & Installation / Re: Install - screen goes blank when pmshell starts
« on: February 05, 2024, 11:05:59 am »
About GPT. It is slightly slower due to the filter. It allows partitions that are aligned correctly, this is important for SSD's etc that use 4KB sectors as ideally the devices sectors and file systems (JFS) should align. Not only somewhat faster but in the case of SSD's, cuts down on wear. Write a one byte file and 4KB gets written. If sectors are not aligned, 8KB actually gets written with house keeping and garbage collection happening twice as often.
With GPT, you can use other partitioning tools and at the worse, have to go back and reassign the drive letter.
You can only boot GPT with UEFI, and you will have to add the GPT filter to older installs to rad the GPT disks
Personally, I have my 1TB NVMe device partitioned GPT with ArcaOS 5.1, a big data partition and Linux on it and my older half a TB SSD partitioned MBR with legacy installs. rEFInd allows me to boot to which ever I choose, with airboot on the MBR device then chain loading.
Edit: ESP is also on the 1TB NVMe device. I did have one on the regular MBR SSD device, which I added to the system after installing 5.1, things got confusing until I wiped that partition. As you found, hiding it is not good enough.

David said at Warpstock (from Arca Noae) the performance impact is 15% slower then an MBR disc. So that is not to bad.
Honestly GPT is MUCH more compatible for setting up multiboot systems.

Roderick

5
Applications / Re: AcpiDaemon start
« on: January 30, 2024, 04:28:44 pm »
I agree with you. No one else has stated that Firefox or seamonkey was crashing for no apparent reason. As stated earlier, it could depend on the hardware.  I have installed ArcaOS multiple times on my laptop, and every time after,  Firefox would start crashing for no apparent reason. I was having to go to windows more often for internet.  It has only been two days, but so far no crashes.  I will continue to see what happens.  I need to see what happens in the next weeks.  I would say this is a tip for anyone that might be having the same problem to try.

The problem is that there can be so many factors on a PC that can trigger something, depending on the speed of the system. Moving one driver in the config.sys can alter memory layout.
Changing some BIOS settings can alter speed.

Roderick

6
Hardware / Re: The last nail in OS/2's coffin
« on: January 30, 2024, 04:25:52 pm »
Has anyone considered that AMD might not blindly follow Intel down this road?  I see this as an attempt by Intel to gain relevance.

This may indeed be a possible scenario, also due to the technical challenges a complete 64 bit transition will create.

That said, I'm fairly sure that the actual x86 architecture will survive for long, before it can be considered a dead horse. Too many installed machines and too much software written for it.
I feel confident OS/2 will survive this generation  8)

Mentore

I share that viewpoint.

Roderick

7
Applications / Re: AcpiDaemon start
« on: January 30, 2024, 09:26:00 am »
What you are saying is that this programmer did not have all his facts straight.  This patch that I applied said it was for Warp version.  So, is this patch ignored after it is applied by Warp and later?

Over the years I have heard more claims that certain patches magically fix something. My take on it is the only to way find out for real is with process dump or a memory dump.
The problem is when the PMSHELL comes up so much code is running and that can trigger other things. A great example was the looping sound example where the startup system sound would loop.
People thought it was a bug in Uniaud. It turned out it was a bug in SND.DLL that deals with the system sounds. David found it while looking at under the kernel debugger.
So yes it MIGHT fix it, but please do not make assumptions to quickly. Moving things around in memory by making buffers bigger, changing load order is not yet an indication its really fixed.
But rather worked around by accident.

Roderick Klein

8
Applications / Re: AcpiDaemon start
« on: January 30, 2024, 08:59:48 am »
If you think there is an issue and you reproduce this on one machine you have with the ArcaOS kernel loaded open a ticket via https://mantis.arcanoae.com.
Attach a testlog and describe how you can reproduce a hang.

Roderick Klein

Hi Roderick

For me, Firefox would always crash and I would have to restart it.  Moving that statement to the startup.cmd and keeping my fingers crossed, it has not crashed in two days.  Before, it was all the time.  On two partitions, not one crash so far.

9
Applications / Re: AcpiDaemon start
« on: January 30, 2024, 01:36:21 am »
If you think there is an issue and you reproduce this on one machine you have with the ArcaOS kernel loaded open a ticket via https://mantis.arcanoae.com.
Attach a testlog and describe how you can reproduce a hang.

Roderick Klein

Hi Roderick

For me, Firefox would always crashed and I would have to restart it.  Moving that statement to the startup.cmd and keeping my fingers crossed, it has not crashed in two days.  Before, it was all the time.  On two partitions, not one crash so far.

Having worked on eCS for 10  years i would be carefull to conclude that certain patches work.
It could but, so much happens in code of the OS st startup. It could be something else.
If you think you can reliably reproduce this issue on Arcaos open a ticket please.
If you discuss it here and its an issue it will never get fixe in Arcaos.

Roderick

10
General Discussion / Re: Hobbes Retirement
« on: January 30, 2024, 12:52:17 am »
Hello Roderick

Nathan is doing a good progress with the replacement development, he has the filesearch links, search inside the "long description" text and other fixes that didn't have the original hobbes. He even coded a web form for file uploads.
Even that I share your same concerns, I say let's see how it works out and see if we can evolve from there.

The important thing to start is being able to backup "Nathan's Hobbes" files (He agreed to that), and keep uploading the backup on different places like OS2Site.com/sw.
After some days of the release we will see if this works or not, and if not, we can ask Nathan if something can be worked out.

But Roderick, I prefer if you keep going and improve OS2Site. I'm also willing to collaborate to have OS2Site as the "Main 'OLD and NEW' Repository". (Like saying HobbesArchive has the new stuff, but OS2Site/sw has everything)
I still dream of reorganizing the OS2Site.com/sw (After a full public backup in archive.org or course). I have on my hard drive (and backup) the same hobbes reorganized structure, but with the difference that I try to collect all older versions of the software. That is 56GB for the moment compared against the 18GB that Hobbes are right now. Yes, it is not "that interesting" to collect older versions of the software for some, but it is worth to be preserved and 56GB is not a lot of storage these days.

Regards

Sorry Martin there is only so many balls I keep in the air at the time (day has 24 hours) . I am willing to work on it but the priority for me has gone down as I simply find the current setup not workable. I also wanted to assemble a team of volunteers to process incoming on os2site.com. I just do not think its useful to have two servers with an upload DIR.

Roderick

11
Applications / Re: AcpiDaemon start
« on: January 30, 2024, 12:48:20 am »
If you think there is an issue and you reproduce this on one machine you have with the ArcaOS kernel loaded open a ticket via https://mantis.arcanoae.com.
Attach a testlog and describe how you can reproduce a hang.

Roderick Klein

12
General Discussion / Re: Hobbes Retirement
« on: January 28, 2024, 02:21:20 pm »
ftp hobbes.os2voice.org is already up for upload.

Thanks Roderick. I just uploaded some old software that Hobbes was missing to give it a try. For me moment my only observation is that you can not see what file were uploaded. "/incoming" always look empty but the files seems to be uploaded without problems.

Regards

I have decided to cancel work on that mirror as Nathan went ahead. As I have stated my concern before. We need one master hobbes archive not two upload FTP.
As we do not know how to synch files.

My personal opinion is that just like with NMSU it would have better for the long run if it would have been ran by OS/2 VOICE or OS/2 world foundation so it does not depend on single person.
I do not think a master archive should be on PC in his son bedroom, sorry. And if something happens to his home (what we do not hope) well we loose it again.
Doing this via an organization provides slightly more redundancy in different ways.

So I see it as a waste of time to work on a mirror as well.

Roderick

13
Hardware / Re: The last nail in OS/2's coffin
« on: January 25, 2024, 01:53:55 pm »
Poked around on other forums, mentioning the X86-S spec, and folks there pointed out some VirtualBox tidbits:

  - old versions of VirtualBox emulated enough of a CPU such that they could run OS/2 when other virtualization products couldn't.
  - newer versions of VirtualBox no longer needed that emulation layer, as they could pass things thru to the CPU directly. But, to this day, VirtualBox still has pieces of QEMU in it.

Most likely, both VirtualBox and QEMU will be able to run OS/2 long after a particular x64-only CPU can't. I'd guess that Intel needs a new line of x64-only CPU's to carry them further into the server ecosystem. But I'd also guess that they'll continue CPU's that support legacy business apps for a long time (lots of money there).

If DOS, an OS that is older than OS/2, can be emulated and survive & thrive to this day, there shouldn't be any reason why OS/2 (all variants) can't be emulated as well, and prosper long into an x64 world. VirtualBox (virtualization) and QEMU (emulation) will get us there.

CPU's might take away a needed "layer of compatibility" on the one hand, but they almost always give back with "speed" on the other hand. Emulation developers do the rest ...

Its not that easy virtualization will full x86 emulation.  It has an impact on perfomance.

Roderick

14
Hardware / Re: The last nail in OS/2's coffin
« on: January 24, 2024, 01:07:51 am »
Correct. Virtualization of OS/2 (and likewise: DOS and Windows 16-bit) will not be possible once x86-S is introduced. Only emulation will be possible (with a huge performance impact, of course).

This raises the question of how virtualization works currently with a 64 bit system running 16 bit code. My understanding is that in 64 bit mode, 16 bit software doesn't work.
As for emulation, with JIT compiler, performance can be pretty good. The PowerPC OS/2 was supposed to run DOS/WinOS2 well, as well as some versions of NT such as the Alpha port.
I doubt that anyone will actually do it but possible.

I do not know how this works, but the CPU does not swtch to legacy mode. You will have to ask a CPU expert.

Roderick

15
Hardware / Re: The last nail in OS/2's coffin
« on: January 23, 2024, 10:20:06 am »
If you want to delve into the details:

https://www.intel.com/content/www/us/en/developer/articles/technical/envisioning-future-simplified-architecture.html

Currently, this is only a proposal. But it gives a very good idea of what Intel thinks is no longer needed.

Its expected this will indeed be long term thing to happen. I understand the legacy bit takes up 12% of the space in CPU.
I read an article (I do not the have link handy) of one of the main AMD tech engineers and he said the proposal was every interesting.
However the engineer indicated the changes proposed are "very, very, very, very complex to make".

The other thing that this would imply is (the way I understand it) is that NO current VM would work anymore
to run legacy OS. Its the CPU that provides the virtualization support to hypervisor.

Roderick

Pages: [1] 2 3 ... 44